Missing samples reconstruction using an efficient and robust instantaneous frequency estimation algorithm

In order to recover missing samples in a nonstationary signal, this paper employs a time-signal analysis and filtering method. The instantaneous frequency of a multicomponent signal is first estimated by employing a robust and computationally efficient method. Then the time-frequency filtering is performed using a dechirping operation to recover missing samples. These steps are repeated until convergence. The proposed method achieves better performance than the state of art methods both in terms of the accuracy of the recovered signal and computational efficiency.